diff options
| author | Henry Jiang <h243jian@uwaterloo.ca> | 2024-06-05 11:35:37 -0400 |
|---|---|---|
| committer | GitHub <noreply@github.com> | 2024-06-05 11:35:37 -0400 |
| commit | b5b61cce96580552a796f21fdcbee3080fd04a9e (patch) | |
| tree | 2b1afd52254d9bcd55d9fbf48811840618992319 /llvm/lib/Bitcode/Writer/BitcodeWriter.cpp | |
| parent | 07b9d231ff9baa6473b0dd588a3ce5330d3e4871 (diff) | |
[VectorCombine] Preserves the maximal legal FPMathFlags during foldShuffleToIdentity (#94295)
The `VectorCombine::foldShuffleToIdentity` does not preserve fast math
flags when folding the shuffle, leading to unexpected vectorized result
and missed optimizations with FMA instructions.
We can conservatively take the maximal legal set of fast math flags
whenever we fold shuffles to identity to enable further optimizations in
the backend.
---------
Co-authored-by: Henry Jiang <henry.jiang1@ibm.com>
Diffstat (limited to 'llvm/lib/Bitcode/Writer/BitcodeWriter.cpp')
0 files changed, 0 insertions, 0 deletions